|
--- |
|
license: other |
|
base_model: apple/mobilevit-small |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- accuracy |
|
model-index: |
|
- name: quickdraw-MobileViT-small-a |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# quickdraw-MobileViT-small-a |
|
|
|
This model is a fine-tuned version of [apple/mobilevit-small](https://huggingface.co/apple/mobilevit-small) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.9705 |
|
- Accuracy: 0.7556 |
|
|
|
## Model description |
|
|
|
More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
More information needed |
|
|
|
## Training and evaluation data |
|
|
|
More information needed |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 0.0008 |
|
- train_batch_size: 512 |
|
- eval_batch_size: 512 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_steps: 5000 |
|
- num_epochs: 8 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | |
|
|:-------------:|:------:|:-----:|:---------------:|:--------:| |
|
| 1.464 | 0.5688 | 5000 | 1.4063 | 0.6493 | |
|
| 1.2318 | 1.1377 | 10000 | 1.2154 | 0.6937 | |
|
| 1.1699 | 1.7065 | 15000 | 1.1495 | 0.7096 | |
|
| 1.1018 | 2.2753 | 20000 | 1.1081 | 0.7190 | |
|
| 1.0837 | 2.8441 | 25000 | 1.0871 | 0.7240 | |
|
| 1.0343 | 3.4130 | 30000 | 1.0550 | 0.7326 | |
|
| 1.0198 | 3.9818 | 35000 | 1.0281 | 0.739 | |
|
| 0.9795 | 4.5506 | 40000 | 1.0125 | 0.7435 | |
|
| 0.9339 | 5.1195 | 45000 | 0.9964 | 0.7475 | |
|
| 0.9292 | 5.6883 | 50000 | 0.9843 | 0.7510 | |
|
| 0.8975 | 6.2571 | 55000 | 0.9795 | 0.7528 | |
|
| 0.8957 | 6.8259 | 60000 | 0.9723 | 0.7548 | |
|
| 0.8721 | 7.3948 | 65000 | 0.9716 | 0.7555 | |
|
| 0.8725 | 7.9636 | 70000 | 0.9705 | 0.7556 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.41.0 |
|
- Pytorch 2.2.1 |
|
- Datasets 2.19.1 |
|
- Tokenizers 0.19.1 |
|
|
